Research map / Label-efficient and robust learning / Recognition and long-tail
Towards Distribution-Agnostic Generalized Category Discovery
Jianhong Bai, Zuozhu Liu, Hualiang Wang, Ruizhe Chen, Lianrui Mu, Xiaomeng Li, Joey Tianyi Zhou, YANG FENG, Jian Wu, Haoji Hu
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| regularize with true (oracle) long-tailed distribution instead of balance prior (ablation) | BaCon pseudo-labeling branch · ImageNet-100-LT | overall accuracy: 58.3 -> 64.6 | Improved |
| estimate distribution from contrastive branch (con k-means) instead of balance prior (ablation) | BaCon · ImageNet-100-LT | overall accuracy: 58.3 -> 62.0 | Improved |
| regularize only with labeled-set distribution instead of estimated distribution (ablation) | BaCon · ImageNet-100-LT | overall accuracy: 62.0 -> 44.8 | Hurt |
| k-means on classifier features instead of contrastive features for distribution estimation (ablation) | BaCon · ImageNet-100-LT | overall accuracy: 62.0 -> 56.4 | Hurt |
| add pseudo-label-based contrastive loss (vanilla) over baseline (ablation) | BaCon · CIFAR-100-LT | overall accuracy: 62.5 -> 65.5 | Improved |
| add debiasing of logits to vanilla pseudo-label loss (ablation) | BaCon · CIFAR-100-LT | overall accuracy: 65.5 -> 66.5 | Improved |
| add sampling of unlabeled instances to vanilla pseudo-label loss (ablation) | BaCon · CIFAR-100-LT | overall accuracy: 65.5 -> 66.6 | Improved |
| combine debiasing and sampling (ablation) | BaCon · CIFAR-100-LT | overall accuracy: 65.5 -> 67.2 | Improved |
| soft pseudo-label contrastive loss instead of hard loss (ablation) | BaCon · CIFAR-100-LT | overall accuracy: 64.9 -> 67.6 | Improved |
| dot product instead of L1/L2/cosine similarity for positiveness score (ablation) | BaCon · CIFAR-100-LT | overall accuracy: 62.7 -> 67.2 | Improved |
| vary re-estimate interval r from 1 to 50 (ablation) | BaCon · CIFAR-100-LT | overall accuracy: 67.9 -> 66.4 | No clear effect |
| vary hyper-parameter k from 0 to 1.5 (ablation) | BaCon · CIFAR-100-LT | overall accuracy: 66.2 -> 67.2 | Mixed |
Closest papers by meaning
- DebGCD: Debiased Learning with Distribution Guidance for Generalized Category Discovery (ICLR 2025 · similarity 0.90)
- Generalized Category Discovery via Reciprocal Learning and Class-Wise Distribution Regularization (ICML 2025 · similarity 0.90)
- Towards Understanding Parametric Generalized Category Discovery on Graphs (ICML 2025 · similarity 0.89)
- Federated Generalized Category Discovery (CVPR 2024 · similarity 0.88)
- Active Generalized Category Discovery (CVPR 2024 · similarity 0.88)
- Solving the Catastrophic Forgetting Problem in Generalized Category Discovery (CVPR 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
